한국생산성학회

As the economy grows, interest in financial investment is increasing. One of the most representative ways of investing in finance is stocks. When investing stocks, the stock price index, which indicates the situation in the market, has a lot of influence on investors' decisions toward effective investment. Until now, traditional time series analysis methods and machine learning methods have been used to perform analyses that predict stock price index data. In this study, a model for analyzing and predicting stock price index data using functional linear methods was presented. By using comparison with functional linear model to present and predict the DTW clustering method implemented using the same data, the efficiency of functional linear model was explained. In the full-scale analysis, the stock price index data were expressed by converting it into functional data through smoothing using linear combinations of base functions, and the functional linear model was implemented and predicted. For comparison, the model was implemented and predicted through DTW clustering. Use DTW methods to classify patterns in data and to identify and predict patterns most similar to those in the forecast period. As a result, you want to compare the accuracy of the two models with MAPE, RMSE, and see the possibility of a functional linear model.